Hyperopt: a Python library for model selection and hyperparameter optimization

By James Bergstra, Brent Komer, Chris Eliasmith, Daniel L. K. Yamins, David D. Cox. 2015

963 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computational Theory and Mathematics

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computational Theory and Mathematics

Citations in the corpus

Optuna: A Next-generation Hyperparameter Optimization Framework (2019, methodology): “Spearmint [2] and GPyOpt use Gaussian Processes, and Hyperopt [1] employs tree-structured Parzen estimator (TPE) [6].”
Gandiva: Introspective Cluster Scheduling for Deep Learning (2018, methodology): “We use Hyperopt to generate 100 configurations, with the search space covering both the neural network architecture and various tunable hyper-parameters.”
Taking Human out of Learning Applications: A Survey on Automated Machine Learning (2018, methodology): “category. Early attempts include [99], [102], [112], [113], which have shown a promising performance of Bayesian optimization for hyper-parameter tuning. Later on, it has been applied in sklearn [7], [114] and Weka [8], [39] for automatic conﬁguration of out-ofbox classiﬁers. More recently, CBO has ”
Incremental on-line learning: A review and comparison of state of the art algorithms (2018, methodology): “All parameter are set by Hyperopt [68] using the Tree-of-Parzen-Estimators [69] search algorithm.”
On Hyperparameter Optimization of Machine Learning Algorithms: Theory and Practice (2020, methodology): “Unlike some of the other libraries that only support a single model, Hyperopt is able to use multiple instances to model hierarchical hyper-parameters.”
Neural Networks for the Prediction of Organic Chemistry Reactions (2016, positive): “Hyperopt 51 in conjunction with Scikit-learn 52 was used to optimize the learning rate, the initial scale, and the ﬁ ngerprint length for each of the molecules.”
Tree-structured Parzen estimator: Understanding its algorithm components and their roles for better empirical performance (2023, background): “Hyperopt: A Python library for optimizing the hyperparameters of machine learning algorithms.”
Predicting electricity consumption for commercial and residential buildings using deep recurrent neural networks (2018, methodology): “Hyperopt is a python library for hyper-parameter optimization, which uses a Sequential Modelbased Optimization (SMBO) method [37, 38].”
Bayesian Optimization is Superior to Random Search for Machine Learning Hyperparameter Tuning: Analysis of the Black-Box Optimization Challenge 2020 (2021, methodology): “The starter kit provided examples using the default conﬁgurations of several different optimization packages: Hyperopt [5], Nevergrad [54], OpenTuner [1], pySOT [14], Scikit-Optimize [30], and TuRBO [15], which serve as baselines.”
Tiresias: A GPU Cluster Manager for Distributed Deep Learning (2019, methodology): “Users can use AutoML [2] to perform this exploration efﬁciently and automatically by using some searching tools [14].”
AutoGluon-Tabular: Robust and Accurate AutoML for Structured Data (2020, background): “Other AutoML Platforms worth mentioning include: autoxgboost (Thomas et al., 2018), GAMA (Gijsbers & Vanschoren, 2019), hyperopt-sklearn (Bergstra et al., 2015), TransmogrifAI (Moore et al., 2018), ML-Plan (Mohr et al., 2018), OBOE (Yang et al., 2018), Auto-Keras (Jin et al., 2019), as well as…”
Simple Spectral Graph Convolution (2021, methodology): “To determine K and α , we used the MetaOpt package Bergstra et al. (2015) with 20 steps to meta-optimize hyperparameters on the validation set of Cora.”
Amazon SageMaker Autopilot: a white box AutoML solution at scale (2020, neutral): “H2O AutoML (Pandey, 2019) (Salesforce, 2018), Auto-Keras (Jin et al., 2019), GAMA (Gijsbers & Van-schoren, 2019) and Hyperopt-sklearn (Bergstra et al., 2015).”
Tabular Data: Deep Learning is Not All You Need (2021, methodology): “To select the model hyperparameters, we used HyperOpt [Bergstra et al., 2015], which uses Bayesian optimization.”
Themis: Fair and Efficient GPU Cluster Scheduling for Machine Learning Workloads (2019, background): “A number of existing frameworks such as Hyperdrive [21] and HyperOpt [4] can intelligently apportion GPU resources between various jobs in a single app, and in some cases also terminate a job early if its progress is not promising.”
Accelerating the Machine Learning Lifecycle with MLflow (2018, methodology): “This can be used to implement multi-step workﬂows or to pass projects a “black box” into automated tools such as hyperparameter search [3].”